Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Report on the ASMS 2016 workshop for the Energy, Petroleum & Biofuels Interest Group : “Petroleum and Biofuels: Handling the Data”
Lateefah Stanford and Mark P. Barrow presiding
Tuesday (June 7th, 2016), 17:45-19:00, Room 301A, level 3 Format:
• Lateefah Stanford (BP) o Welcome o Introduction to workshop
• Mark Barrow (University of Warwick, United Kingdom) o Overview of available petroleomics software o Examples of approaches used by different groups
• Ralf Zimmermann (University of Rostock and Helmholtz Zentrum München, Germany) o GCxGC TOF MS, thermogravimetry photoionization MS, GC-APCI-FTICR MS o SPI and REMPI; aliphatic and aromatic species o Variety of data analysis and visualization methods for different data o Questions/Discussion
• Sunghwan Kim (Kyungpook National University, South Korea) o Petroinformatics: combining petroleomic data and statistical tools to predict physical and
chemical properties o Use of multiple methods (MS, LC, IMMS, NMR, IR); need to combine data o Proposal for discussion: conversion of DBE information to structures, to be compared with NMR o Questions/Discussion
At the beginning of the workshop, over 50 attendees were counted, although more also arrived later. The attendees represented both academic and commercial establishments, and the workshop lasted the full duration of the allocated time. Lateefah Stanford opened the workshop, providing the welcome and introduction. Mark Barrow provided a short presentation, with an overview of different petroleomics software tools available, examples of types of data being produced by different laboratories, and examples of possible topics for discussion, where the workshop should be discussion-led. Two speakers were invited to highlight potential areas for further discussion. Ralf Zimmermann presented data produced using a range of experimental methods. GCxGC TOF MS data were displayed using multiple dimensions, including m/z or Kendrick mass defect. For thermogravimetry measurements, temperature vs. m/z was plotted and comparisons of ionization methods were made by color coding data points according to whether peaks were observed only by single photon ionization (SPI), resonance-enhanced multiphoton ionization (REMPI), or both methods. Thermal analysis or GC was coupled with FTICR MS, where the data was plotted as retention time or temperature vs. m/z. When GC was used, artifact species from the ionization process could be excluded by ensuring correlation of peaks of interest with appropriate retention times. Sunghwan Kim discussed the concept of “petroinformatics,” using statistical tools to ultimately develop a predictive system for the physical and chemical properties of petroleum. FTICR MS, Orbitrap, 2D GC, LC, and NMR data was shown and there was a question of how to appropriately combine the results from different experiments. Sunghwan presented an idea for combining DBE information with NMR data to determine structures of the components, based upon assumed “basic structures” for different compound classes. The total number of aromatic and non-aromatic carbon atoms could then be calculated, and the calculations were compared with NMR data. The audience was asked to discuss this proposal. It was clear that there is an increasing diversity of experimental methods for the characterization of petroleum and biofuels, and there is a need to develop better methods for combining and understanding the data. Adding further dimensions to experiments, such as temperature or time, also leads to the acquisition of larger data sets and so creates greater challenges for storage, compression, and analysis. As the outgoing interest group coordinator, Lateefah called for people to put their names forwards if they are interested in being one of the interest group coordinators for next year. By the end of June, one person had contacted the coordinators with an interest in the role.
Petroleum and Biofuels: Handling the Data
Energy, Petroleum & Biofuels Interest Group
Lateefah Stanford (BP)
Mark P. Barrow (University of Warwick, UK)
Categorization
1. “Heteroatom class” or “Compound class”
2. Carbon number
3. Double bond equivalents (DBE)
DBE = 1+ c – h/2 + n/2
Energy Fuels 2006, 20, 1664-1673
Software
1. Composer Sierra Analyticshttp://www.massspec.com/
2. PetroOrgFlorida State Universityhttp://software.petroorg.com
3. mzCruiser for Petroleomics Manhoi Hurhttps://github.com/mhhur/Petroleomics
Ionization methods
Rapid Commun. Mass Spectrom. 2014, 28, 1345-1352
Chromatography
Anal. Chem. 2014, 86, 8281-8288
Gas Chromatography Coupled to Atmospheric Pressure ChemicalIonization FT-ICR Mass Spectrometry for Improvement of DataReliabilityTheo Schwemer,†,‡ Christopher P. Ruger,† Martin Sklorz,*,†,§ and Ralf Zimmermann†,‡,§
†Joint Mass Spectrometry Centre/Chair of Analytical Chemistry, University of Rostock, 18051 Rostock, Germany‡HICE − Helmholtz Virtual Institute of Complex Molecular Systems in Environmental Health − Aerosols and Health, 85764Neuherberg, Germany, www.hice-vi.eu§Joint Mass Spectrometry Centre/Cooperation Group Comprehensive Molecular Analytics, Helmholtz Zentrum Munchen, 85764Neuherberg, Germany
*S Supporting Information
ABSTRACT: Atmospheric pressure chemical ionization (APCI) offersthe advantage of molecular ion information with low fragmentation.Hyphenating APCI to gas chromatography (GC) and ultrahigh resolutionmass spectrometry (FT-ICR MS) enables an improved characterization ofcomplex mixtures. Data amounts acquired by this system are very huge,and existing peak picking algorithms are usually extremely time-consuming, if both gas chromatographic and ultrahigh resolution massspectrometric data are concerned. Therefore, automatic routines aredeveloped that are capable of handling these data sets and further allowthe identification and removal of known ionization artifacts (e.g., water-and oxygen-adducts, demethylation, dehydrogenation, and decarboxyla-tion). Furthermore, the data quality is enhanced by the prediction of an estimated retention index, which is calculated simplyfrom exact mass data combined with a double bond equivalent correction. This retention index is used to identify mismatchedelemental compositions. The approach was successfully tested for analysis of semivolatile components in heavy fuel oil and dieselfuel as well as primary combustion particles emitted by a ship diesel research engine. As a result, 10−28% of the detectedcompounds, mainly low abundant species, classically assigned by using only the mass spectrometric information, were identifiedas not valid and removed. Although GC separation is limited by the slow acquisition rate of the FT-ICR MS (<1 Hz), a databasedriven retention time comparison, as commonly used for low resolution GC/MS, can be applied for revealing isomericinformation.
Complex analytical matrices requires improved sensitivityand resolution of instruments to provide a more accurate
and comprehensive characterization of investigated samples.Ultrahigh resolution mass spectrometry (HR-MS) can feasiblybe used to separate and assign elemental compositions to morethan 17 000 species in a complex sample1 within a few minutes.But without preseparation or application of MSn, it is possibleneither to distinguish between analytes and fragments nor todifferentiate isomeric species without further fragmentation. Ionsources applied for HR-MS such as electrospray ionization(ESI), atmospheric pressure chemical ionization (APCI), oratmospheric pressure photo ionization (APPI) mainly formmolecular ions, which facilitate the mass spectra and datainterpretation. Unfortunately, these techniques are not devoidof ionization artifacts2,3 and competition,4 leading to falseassignments and reduced reliability.Since the first successful coupling of gas chromatography
(GC) to a FT-ICR MS,5 no significant advantages over theconventional GC/MS systems with electron ionization (EI)have been demonstrated. The hyphenation of GC and APCI to
Fourier transform ion cyclotron resonance mass spectrometry(FT-ICR MS), which was first introduced by Barrow et al. in2014,6 could change this situation for certain application areas.In contrast to the common GC/MS systems, this techniqueprovides ultrahigh resolved mass spectrometric information formolecular ions, but a huge amount of data is acquired (severaltens of gigabytes) and data analysis is extremely time-consuming. As a consequence, Barrow et al. reduced the databy averaging mass spectra for a 2 min time interval. Thisfacilitates the data analysis but does not exploit the whole of thechromatographic information. Generally, the assignment ofelemental composition to exact masses is restricted byinstrumental mass accuracy and resolution.7,8 Even theapplication of data analysis and reduction routines9−11 forfiltering chemically correct elemental compositions is notalways sufficient to provide unique assignments. The accuracy
Received: June 4, 2015Accepted: November 11, 2015
Technical Note
pubs.acs.org/ac
© XXXX American Chemical Society A DOI: 10.1021/acs.analchem.5b02114Anal. Chem. XXXX, XXX, XXX−XXX
Anal. Chem. 2015, 87, 11957-11961
Ion mobility mass spectrometry (IMMS)
Anal. Chem. 2009, 81, 9941-9947
the sample which created unresolved interferences and lessaccurate mass measurements when using a QToF MS com-pared with the very high mass resolution achievable with anFT MS. The data were then reviewed to evaluate how the
limitations of comparatively lower mass resolution may beovercome by inclusion of the ion mobility data points. It wasobserved that species that belong to the same heteroatomfamily do not only have sequences of recognisable mass
Fig. 6 Kendrick plot of anitrogen-containing group inthe resin sample generated fromFT MS data
Fig. 7 A plot of drift time versus m/z showing three different DBE series for one particular nitrogen-containing family. Consideration of the ionmobility data helped to isolate an incorrectly assigned ion when selection was based on mass-to-charge ratio alone
100 Int. J. Ion Mobil. Spec. (2013) 16:95–103
Int. J. Ion Mob. Spect. 2013, 16, 95-103
Statistical analysis
Mass Spectrom. Rev. 2014, 34, 248-263
Speakers
• Ralf Zimmermann University of Rostock and Helmholtz Zentrum München, Germany
• Sunghwan KimKyungpook National University, Republic of Korea
Example topics
• Data analysis and visualization
• Additional dimension to data (e.g. temperature, time, etc.);
hyphenated techniques
• Large data sets
• Comparison of different instrumentation and methods
• Statistics, chemometrics
Workshop will be discussion-led
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Ralf Zimmermann, Martin Sklorz, Theo Schwemer, Christopher Rüger, Thorsten Streibel, Mohammad Saraji-Bozorgzad, Andreas Walte, Thomas Gröger
Joint Mass Spectrometry Centre University of Rostock, Chair of Analytical Chemistry, Inst. of Chemistry, Germany and Helmholtz-Zentrum
München, CMA, Germany ([email protected])
Photonion GmbH Schwerin, Germany
Multiple novel techniques for petrochemical analyses … … require multiple data analysis approaches
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Dalluge
Massspectrometry
++HV -HV -HV
MCP
TOFMS
M
GCoven
1.GC
2.GC
Modulator
EI
Detector
Dallüge et al.
Mass (m/z)
Addressingcomplexpetrochemicalmolecularmixtures:EnhancedselecHvitybychromatographicseparaHonèComprehensivegaschromatography(GCxGC)-TOFmassspectrometry
Petrochemical Analyses: Resolution, resolution, resolution … (mass, chromatographic, process time … )
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
GCxGC chromatogram plot: è Visualizes enhanced GC separation
LDI-FTMS
Kendrick mass defect plot (PAH class) è Visualizes enhanced mass separation
Comprehensive high resolution 2D GC (GCxGC) and high resolution mass TOF spectrometry
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Accurate mass information could be used to calculate elemental composition and Kendrick mass defects from verified M+ peaks
RetenHonTimeFirstDimension
RetenH
onTim
eSecond
Dim
ensio
n
50 450
KMD
KM(Da)
PureHydrocarbons Oneormoresulfur Oneormorenitrogen
Analysis of the aromatic composition of heavy fuel oil by GCxGC-HRTOFMS
AromaHcCompounds
LECO GCxGC-HRTOF (Pegasus HRT, R ~ 38.000, < 1 ppm)
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
KMD
Mass defect information could be also be integrated with the 2 D chromatographic data in 3D representation
SecondDimension
FirstDimension
KMD
KMD
Analysis of the aromatic composition of heavy fuel oil by GCxGC-HRTOFMS
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Thermogravimetry - Photo Ionization-Mass Spectrometry: Including soft ionisation MS and thermal process resolution in the analysis
Single photon ionization (SPI) with VUV photons èSoft ionization of aromatics and saturates TG - PIMS
Saraji et al., Anal. Chem. 80 (2008) 3393 Geißler et al., Anal. Chem. 81 (2009) 6038
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
http://www.scribd.com/doc/51719364/ CRUDE-CHARACTERIZATION
After Mietek Boduszynski
TG-SPI-MS and TG-REMPI-MS (selective “aromatic” ionization) results plotted together
Thermogravimetry - Photo Ionization-Mass Spectrometry: Including soft ionisation MS and thermal process resolution in the analysis
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Crude oil 1 Crude oil 2
SPI REMPI overlapp
Thermogravimetry - Photo Ionization-Mass Spectrometry: Including soft ionisation MS and thermal process resolution in the analysis
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
• APCI-FTICR easily switchable between GC- and TG-mode
Thermal analysis (TA) or gas chromatography (GC) coupled to APCI-FTICR: Adding the time dimension to UHR-MS
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Retention time
Gas chromatography (GC) coupled to APCI-FTICR: Adding the GC retention time dimension to UHR-MS
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Comparison of thermal analysis- and gas chromatographic separation-APCI-FTICR results (Heavy Fuel Oil)
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
birch bark pellets
oak
Thermal analysis (TA) coupled to APCI-FTICR: Adding the process time/temperture dimension to UHR-MS
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Thermal analysis (TA) coupled to APCI-FTICR: Adding the process time/temperture dimension to UHR-MS
R. Zimmermann et al. ASMS 2016, San Antonio, TX, USA, June 5th-9th 2016
Acknowledgements Zimmermann Group @ Joint Mass Spectrometry Centre Dr. J. Maguhn Dr. T. Streibel Dr. J. Schnelle-Kreis Dr. M. Sklorz Dr. J. Lintelmann Dr. E. Feicht Dr. M. Kirchner Dr. J. Passig Dr. G. Jakobi Dr. S. Smita Dr. J. Orasche Dr. G Ferron Dr. S. Haack Dr. S. Ehlert Dr.. S. Öder E. Hübner T. Gröger T. Miersch A. Sutherland C. Rüger T. Schwemer G. Dragan H. Czech V. Kohlmeier J. Sanchez
S. Erdmann B. Weggler P. Sattler G. Abbazade M. Oster E. Karg B. Gruber M. Jennerwein F. Li L. Müller A. Rheda X. Wu L. Müller K. Lau M. Hahn Y. Huang J.Varga S. Scholtes C. Kühl R. Reiss A. Hoffmann A. Ulbrich C. Grimmer
Joint Mass Spectrometry Centre Funding: • Uni Rostock and State of Mecklenburg Vorpommern • Helmholtz Zentrum München • Deutsche Forschungsgemeinschaft (DFG) • Bundesministerien für Bildung &Forschung/Wirtschaft (BMBF, BWM) • Bayerische Forschungsstiftung (BFS) • Deutsche Gesetzl. Unfallversich. (DGUV) • Bundeskriminalamt (BKA) • Companies (SASOL Ltd., Netzsch GmbH, Photonion GmbH, LECO, Airsense GmbH, Borgwaldt KC, SABIC Inc., Shimadzu etc.) • Helmholtz-Impulse and Networking- Fonds (Virtual Helmholtz Institute - HICE)
M a ss sp ec tro m etry L a b .
Combining Data and
Data Mining in Petroleomics
EEMSLab Kyungpook National University
Department of Chemistry
Environ. & Energy Mass spec. Research Lab
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
Statistical tools
Petroleomic + other data
Prediction system
Petroinformatics
“It should be possible to predict the chemical and physical properties of crude oils from the chemical compositions determined using FT-ICR MS.”
M a ss sp ec tro m etry L a b .
m/z700650600550500450400350
m/z469.279469.238469.197469.156469.115469.074469.033
12 3
4
5
6
7
8
9
10
11
1213
14
15
16
17
18
Finding a needle in a haystack!!!
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
Cru
de o
il Saturate
Aromatics
Resins
Asphaltenes
GC 2D-GC
2D-GC, LC HR-MS, NMR IM-MS, IR
HR-MS IM-MS
NMR, IR
HR-MS, NMR IM-MS
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
Connection between properties and compositions
Energy Fuels 2016, 30, 915−923
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
M a ss sp ec tro m etry L a b .
DBE vs #C plot as a basis for combining data
Environ. & Energy Mass spec. Research Lab
M a ss sp ec tro m etry L a b .
[QHD aromatics] [APM-FA aromatics] (a) Orbitrap MS
(b) FT-ICR MS 2,0001,5001,0005000
50
0
-50
2,0001,5001,0005000
50
0
-50
Rel
ativ
e si
gnal�
0�
-100�
100�
Time (sec)� 1� 2�
300� 1000�100� 600� 300�100� 1000�600�
Rel
ativ
e si
gnal�
0�
-100�
100�
Time (sec)� 1� 2�
Environ. & Energy Mass spec. Research Lab
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
M a ss sp ec tro m etry L a b .
Chrysene Figure S1.
M a ss sp ec tro m etry L a b .
(a) HC class
(8, 5)
(10, 7)
(10, 5)
[APM-FA aromatics]
DB
E
Carbon number 10 20 30 40 50 60 70 80
0
10
20
30
40
50
(8, 3)
(12, 8) (13, 8)
(12, 6) (11, 6) (13, 6)
(12, 7) (13, 7) (14, 7)
[QHD aromatics]
(18, 13) (19, 14)
(16, 12)
(10, 7)
DB
E
FT-ICR MS Orbitrap MS Degree of difference (log2FC)
Overlap area
Degree of difference (log2FC)
0
10
20
30
40
50
10 20 30 40 50 60 70 80
(16, 8) (17, 8)
(15, 9) (16, 9) (17, 9)
(16, 10) (17, 10) (18, 10)
Manuscript in preparation
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
Converting DBE to structures and c
omparison with NMR data
Manuscript in preparation
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
Prep-scale separation
a flow rate of 25 mL/min 0.1~0.3 g of sample loaded
Figure 1. MPLC-ELSD spectrum of an AR maltene.
Energy & Fuels, in press
Elution time(min)
ELSD
, Int
ensi
ty(m
V) Saturate Aro1 Aro2 Polar1 Polar2
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
1H NMR %1Hnon-aro
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
Converting DBE to # of aromatic carbons
1. All the compounds with DBE values smaller than the ones of the basic
structures are considered non-aromatic.
2. Linear polyaromatic hydrocarbon structure is dominant. Therefore, an
increase of 3 in the DBE value is equivalent to addition of an aromatic
ring to the basic structure.
3. Increase of 1 and 2 in the DBE value from the aromatic structured obt
ained from assumption 1 and 2 is not caused by increased number of
aromatic ring.
Manuscript Submitted
u Basic structures
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
DBE Total #C Aromatic #C %Cnon-aro
7 15 10 (15-10)/15 8 16 10 (16-10)/16
12 20 12 (20-12)/20
HC class
DBE=4 DBE=7 DBE=10
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
DBE à # aromatic ring in a given formulae à # aromatic carbon in the formulae
à (total # of carbon - # aromatic carbon) = total # of carbon %
Cno
n-ar
o
%1Hnon-aro
R² = 0.82
0
10
20
30
40
50
60
70
80
82 84 86 88 90 92 94 96 98 100
Manuscript Submitted
M a ss sp ec tro m etry L a b .Environ. & Energy Mass spec. Research Lab
l Development of a good algorithm and sub
sequent development of software
l Any comment is welcome!!!